Founders ask Collective 54 this once in our records, and not in 2026. The churn rate, ideal client and pricing pushback answers on this site cover retention targets, fit and price objections; this page covers the decision to bend your own rules to keep a client.
The client retention essay in the newer book says that once revenue becomes recurring, retention becomes the constraint. Below roughly 90 percent retention, it says, new revenue replaces lost revenue instead of compounding, expansion feels uphill and forecasts feel optimistic until they are not. Above it, growth becomes calmer and the firm starts to look like a transferable asset. A founder who has absorbed that message will feel every lost client as a step backward.
The fee quality chapter of the 2020 book pushes the other way. It says most boutiques mistakenly think all revenue is good revenue, that some fees are more valuable than others, and that bad fees decrease the value of the firm and can prevent a sale. As an inference, both are right. The retention line is about keeping the clients you want; it is not an argument for keeping every client at any cost.
As an inference, the answer depends almost entirely on the reason the client is on the way out, so name it before deciding anything. The published sources point to several common reasons.
The client does not fit. The ideal client answer on this site covers defining who you serve; a client outside that profile pulls delivery, pricing and attention away from the work you want to be known for.
The client does not make money. The account management essay lists account-level costing and showing which clients create or destroy EBITDA among the work AI can now do. The engagement manager essay defines contribution margin as fees collected minus direct labor, delivery tools, AI costs and subcontractors, and calls it the primary scoreboard.
The client is a risk. The client relationships chapter of the 2020 book says buyers examine client quality and the health of the end clients a firm depends on, and that no single client should exceed 10 percent of billings.
The client will not respect scope. The engagement manager essay says great engagement managers treat scope creep as an economic decision rather than client service, and offer options: reduce scope, extend the timeline or increase the budget.
Each reason calls for a different response. An unprofitable client may be fixable through price or scope. A client outside your profile usually is not.
As an inference, there are a few situations where bending the rule is a reasonable business decision. The client is a reference or a door into a segment you are entering, and the value of that is real and specific. The client is going through a temporary disruption, such as a leadership change, and the relationship is likely to return to normal terms. Or the departure would push another client over the concentration line or leave a team without work, and you need time to replace the revenue.
The client retention essay says retainers usually fail from perceived stagnation rather than dissatisfaction, and fractional roles from drift as leadership teams change. As an inference, if the client is leaving for one of those reasons, the better move is often to re-anchor the value of the work rather than to change your terms.
The pricing essay in the newer book says that when pricing is not governed, efficiency leaks through discounting, scope expansion and packaging confusion, and one-time exceptions become invisible norms. It says governance includes enforcing price integrity, controlling discounting, managing packaging creep and connecting quoted price to realized margin, and that humans keep the intentional exceptions while the system tracks them.
As an inference, that gives you the rule. If you make an exception, write it down: what you are changing, why, what it costs, who approved it and when it ends. Review it at that date. An exception no one can see is not an exception; it is your new price.
The competitor chapter of the 2020 book warns against discounting so far that it signals you are cheap. The discount answer on this site recommends that any concession be visible and time-limited. As an inference, a client who stays only because the price fell has told you what they think the work is worth, and the discount tends to become the starting point at renewal. If you must give something, give scope or timing flexibility that is easier to reverse.
The engagement manager essay says the growth and retention of the team an engagement manager leads is a direct measure of how well they lead, and lists protecting the team from unnecessary pressure among the moments that cannot be handed to AI. As an inference, a client you would normally let go is often one the team already finds hard to serve. Keeping them by exception asks your people to absorb the cost. Ask the team before you decide, and if you keep the client, tell the team what will change and when the exception ends.
As an inference, retention bought with exceptions is weaker than it looks. A buyer of the firm, the exit essay from Collective 54 says, is judging how much durable profit will survive the transfer. If a share of your retained clients stay only on special terms, your retention rate and your margin tell different stories. Track exception clients separately so you know which retention you earned.
The client relationships chapter tells the SBI story of sales leaders who left their companies and hired the firm again at the next one. As an inference, how you end a relationship matters for that reason. Give notice, finish what you committed to, hand over cleanly and, where it helps, suggest another firm better suited to the work. A client released with care can still refer you.
Collective 54 publishes no policy on when to release a client or which exceptions to allow. The published positions are the roughly 90 percent retention line, retention failure modes by revenue type, not all revenue being good revenue, client quality and concentration, contribution margin as the scoreboard, scope creep as an economic decision, governed pricing and tracked exceptions, the warning against looking cheap, team development as a measure of leadership, and durable profit at exit.
If losing the client would put the firm in financial danger, as an inference, keep them while you replace the revenue, with a dated plan to do so.
If the client is new and the problems are teething issues, fix the engagement before judging the relationship.
And if you find yourself making the same exception for several clients, it is not an exception; change the standard or change the offer.
Usually no. Retention matters, but the 2020 book is clear that not all revenue is good revenue. Name why you would release the client, and make an exception only for a specific strategic reason, in writing, priced, approved, time-limited and reviewed. Avoid deep discounts, ask the team, track exception clients separately in your retention numbers, and if you do let a client go, end the relationship with enough care that they would still refer you.
Collective 54 sets no rule. The 2020 book says not all revenue is good revenue and that bad fees reduce the value of the firm. As an inference, release clients who do not fit or do not make money.
The pricing essay says one-time exceptions become invisible norms without governance. As an inference, write each one down with a reason, an approver and an end date, and review it.
As an inference, only briefly. The retention essay sets a roughly 90 percent line, but the engagement manager essay makes contribution margin the primary scoreboard.
As an inference, give notice, finish your commitments, hand over cleanly and suggest a better fit where you can. The 2020 book shows how departing client contacts can hire a firm again.
Sources: Greg Alexander, The AI-Native Boutique Firm (Advantage Books, January 2027), specifically The AI Client Retention Manager for retention as the constraint in recurring revenue firms, the roughly 90 percent threshold and failure modes by revenue type including stagnation in retainers and drift in fractional roles; The AI Pricing Manager for leakage through discounting, scope expansion and packaging confusion, one-time exceptions becoming invisible norms, governance and humans keeping intentional exceptions; The AI Account Manager for account-level costing and clients that create or destroy EBITDA; The AI Engagement Manager for contribution margin as the primary scoreboard, scope creep as an economic decision with options, team growth as a measure of leadership and protecting the team. Greg Alexander, The Boutique: How to Start, Scale, and Sell a Professional Services Firm (Advantage, 2020), chapter 32 for not all revenue being good revenue and bad fees pushing buyers away; chapter 31 for client quality, healthy end clients, the 10 percent concentration limit and the SBI story of departing sales leaders hiring the firm again; chapter 3 for not discounting so far that you look cheap. Greg Alexander, Why Some Boutique Firms Exit Cleanly and Others Never Really Do (Collective 54), for durable profit at exit. Related Collective 54 answers on this site: what churn rate should I expect, and how is retention measured; who is our ideal client and how do we define and target our ICP; should I discount, and how do I avoid sticker shock later; how do I manage scope changes without letting them blow the budget; how do I handle client pushback on my pricing. Note on scope: Collective 54 publishes no client release policy. Naming the reason first, the situations where exceptions make sense, the written exception with an end date, preferring scope over price concessions, asking the team, tracking exception clients separately, ending relationships well and the flips are inferences used here to organize the source material rather than published Collective 54 positions.
Collective 54 is the private community for founders and executives of boutique professional services firms between $5M and $50M in revenue. Members work these answers against their own numbers.